2D Contour Optimization: Particles flipping and self-intersecting during splitting and PCA

1

Hi,

I am currently working with 2D femur contours, but I am still struggling to get the particles to cover 100% of the shape across all my patients. There are still some gaps where the particles just don’t want to spread.

I haven’t moved on to Phase 2 (correspondence optimization) yet, as I understand I need to achieve perfect coverage on all femurs first. Because of this lack of full coverage and correspondence, the “Overall Mean” in the Analyze tab looks like a messy jumble of overlapping points, and when adjusting the standard deviation slider in the PCA analysis, the points seem to invert, cross over each other and the shape collapses (self-intersection).

I have also tried restarting the process from scratch, closely monitoring the particles at every step to ensure the correspondence remains valid and no points cross.

However, even with a very gradual approach, the particles start to cross and invert before they have even finished covering the entire femur contour. It seems that as the number of particles increases during the splitting phases, the repulsion forces or the regularization cause them to ‘jump’ across the shape’s neck or narrow areas instead of following the 2D boundary.

fb54def3c4b99bdec625c54af1a5d83d1d070448_2_1035x603
fb54def3c4b99bdec625c54af1a5d83d1d070448_2_1035x6031035×603 60.6 KB

I will attach a screenshot of my Optimization screen. Here are the parameters I am currently using:

  • number_of_particles: 512

  • use_normals: 0

  • normals_strength: 0.0

  • iterations_per_split: 5000

  • optimization_iterations: 3000

  • starting_regularization: 100 (I lowered this to give the system more flexibility, but gaps remain)

  • ending_regularization: 10

  • initial_relative_weighting: 0.0001

  • relative_weighting: 100

  • procrustes: 1

  • procrustes_interval: 1

  • procrustes_scaling: 1

  • procrustes_rotation_translation: 1

  • verbosity: 0

  • multiscale: 0

  • multiscale_particles: 0

  • narrow_band: 250

  • use_geodesic_distance: 1

33ed50daed64546171841f83d88fad8f00ba0892_2_1035x586
33ed50daed64546171841f83d88fad8f00ba0892_2_1035x5861035×586 83.8 KB

My questions are:

  1. What parameters would you recommend tweaking to force the particles to cover the remaining gaps without breaking the system?

  2. Is there a specific parameter to prevent self-intersection in 2D closed loops during PCA?

  3. Would placing initial landmarks help guide the particles to cover the whole femur?

Thank you very much!